jamesstringer90 / jamesstringer90/appsandbox
AMD Radeon 780M, GPU acceleration not working in AppSandbox v0.1.4
Nobody has claimed this yet.
- Dominant language
- C
- Stars
- 704
- Forks
- 75
- Avg merge
- 9m
- Merged PRs (30d)
- 6
Description
#Hardware
- Mini PC: AZW SER8
- CPU: AMD Ryzen 7 8845HS
- GPU: AMD Radeon 780M integrated GPU
- OS: Windows 11 pro
- Frigate: Docker
- AppSandbox v0.1.4
Problem
I am trying to use the AMD Radeon 780M integrated GPU for hardware video decoding in Frigate.
According to the Frigate documentation, modern AMD GPUs are supported through VAAPI using the radeonsi driver and the preset-vaapi configuration.
I configured:
ffmpeg:
hwaccel_args: preset-vaapi
and:
LIBVA_DRIVER_NAME: radeonsi
However, hardware video acceleration does not work correctly in AppSandbox.
With hardware acceleration disabled, the camera works correctly using software decoding.
The problem can be reproduced with a single camera, so this does not appear to be related to the number of cameras or the overall Frigate workload.
Expected behavior
The Radeon 780M should be detected and used by FFmpeg/Frigate for VAAPI hardware video decoding.
I would expect the corresponding DRM render device, for example:
/dev/dri/renderD128
to be available and usable by Frigate.
Actual behavior
When using:
ffmpeg:
hwaccel_args: preset-vaapi
hardware video decoding fails.
When I remove the hardware acceleration configuration, the same camera stream works correctly with CPU/software decoding.
Therefore, the issue appears to specifically concern the VAAPI hardware acceleration path with the Radeon 780M.
Question
Is the AMD Radeon 780M fully supported for VAAPI hardware video decoding in Frigate when running inside AppSandbox?
If it is supported, what is the recommended configuration to make the Radeon 780M available to FFmpeg/VAAPI?
In particular, I would like to know whether there are any specific requirements for:
/dev/dri/renderD128radeonsi- Mesa VAAPI drivers
- FFmpeg VAAPI
- Radeon 780M / RDNA3
Relevant Frigate configuration
ffmpeg:
hwaccel_args: preset-vaapi
Environment:
LIBVA_DRIVER_NAME: radeonsi
Additional information
The Radeon 780M is the integrated GPU of the AMD Ryzen 7 8845HS.
I am only testing hardware video decoding with one camera stream.
Software decoding works correctly, but hardware decoding with GPU does not.
Could you please advise whether this is a known limitation or whether additional configuration is required for the Radeon 780M?
Thank you.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with the AppSandbox environment and the Frigate ffmpeg hwaccel_args/preset-vaapi configuration; check whether /dev/dri/renderD128 is exposed and usable with LIBVA_DRIVER_NAME=radeonsi. Done means determining whether the Radeon 780M VAAPI path is supported in AppSandbox and documenting the required configuration or the limitation.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- docker
- Domain
- infrastructure
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 32/100